Prior to the introduction of s23J into the South African Income Tax Act, many specific allowance sections in our Tax Act include anti-avoidance provisions directed against transfers between “connected persons” (i.e. s11(e), s11B, s11D, s11(gA), s11(gC), s12B, s12C, s12G, s14 and s14bis). These provisions were necessary to ensure that appreciated capital assets (e.g. certain plant, equipment and intellectual property) are not re-valued and transferred to another company within the same group in order to “roll-over” the inflated allowance. For example, s11(gC)(bb) (allowances for the acquisition of intellectual property (IP)) provided that:
(bb) where any such invention, patent, design, copyright or other property or knowledge was acquired from any person who is a connected person in relation to the taxpayer, the allowance under this paragraph shall be calculated on an amount not exceeding the lesser of the cost to that connected person of acquiring, devising, developing or creating that invention, patent, design, copyright or other property or knowledge or the market value of that invention, patent, design, copyright or other property or knowledge as determined on the date upon which it was acquired by the taxpayer.
Let’s consider a group of companies comprising ParentCo Ltd, and wholly owned subsidiaries CompanyA (Pty) Ltd, CompanyB (Pty) Ltd, and CompanyC (Pty) Ltd. Now assume that CompanyA initially developed IP in 1980 for R1m and that the IP was subsequently revalued at R200m. Should CompanyB acquire that IP by assignment from CompanyA for R200m, operation of s11(gC)(bb) acted to limit CompanyB’s allowance to the lesser of:
(a) the cost to CompanyA to develop the IP = R1m; or
(b) the market value = R200m (assuming, of course, that the valuation is legitimate)
Since, in many instances, the previous IP related allowances claimed by CompanyA would be recouped, no net tax benefit would be enjoyed by the group as a consequence of concluding this transaction. In fact, the group would be better off relying on the Corporate Rules in our Income Tax Act.
In an attempt to circumvent this anti-avoidance provision, taxpayers frequently interpose an intermediary between CompanyA and CompanyB. The intermediary could either be an unconnected entity (such as a bank) or a fellow group company, such as CompanyC. The interposition of CompanyC was intended to divide the assignment from CompanyA to CompanyB into two parts: a first assignment from CompanyA to CompanyC and a second assignment from CompanyC to CompanyB. Often, the purchase price was merely credited to the purchaser’s loan account. This arrangement was thought to have the following tax effect:
Assignment from CompanyA to CompanyC
S11(gC) allowance that may be claimed by CompanyC is limited by ss(bb) to the lesser of:
(a) the cost of the IP to CompanyA = R1m; or
(b) the market value of the IP = R200m
Assignment from CompanyC to CompanyB
S11(gC) allowance that may be claimed by CompanyB is limited by ss(bb) to the lesser of:
(a) the cost of the IP to CompanyC = R200m; or
(b) the market value of the IP = R200m
That was easy!
Well, maybe not. Anti-avoidance provisions are inserted into our Income Tax Act to prevent specific transactions. When these “sniper” provisions are “easily circumvented” it is generally because some principle of law has been overlooked.
The above transaction can be challenged on three grounds: that CompanyC never obtained title in the IP to assign to CompanyB; that the assignment to CompanyC can effectively be ignored or explained away; and that the whole transaction comprises a single “composite transaction” in terms of which the IP is assigned by CompanyA (the assignor) to CompanyB (the assignee) (albeit via CompanyC).
Ignoring or explaining away the assignment to CompanyC
An immediate on-sale of an asset is a strong indicator that transfer of ownership to the original purchaser (in this instance, CompanyC) was not intended. There have been numerous cases dealing with disguised pledge transactions in which simultaneous “sale” and “on-sale” of assets were held to be disguised (in this regard, see S v Dorfler 1971 (4) SA 374 (AD); Vasco Dry Cleaners v Twycross 1979 (1) 603 (AD); and Bank Windhoek Bpk v Rajie and another 1994 (1) SA 115). Basically, where no satisfactory explanation is provided for the on-sale, the initial “sale” is frequently exposed as a disguised transaction. For example, see McAdams v Fiander’s Trustee & Bell 1919 AD 207 at p230:
“But it is an element to be taken into consideration that he buys in order to sell at once and purports to resell at the same figure plus interest. It is clear that the cattle were merely pledged to Mack in return for lending his name to Fiander & Co. on the bill.”;
Also see S v Dorfler 1971 (4) SA 374 (AD) at 377D-E, where the fact that the purchaser simultaneously bound himself to re-sell at the same price to the so-called seller and that the goods remained in the possession of the complainants throughout was sufficient to expose the true nature of the “sale” agreement:
“One starts off at the outset with the situation that one of the basic essentials of a genuine contract of sale, i.e. the intention on the part of the purchaser to acquire ownership and to deal with the property as owners and the corresponding intention on the part of the seller to confer that ownership on him, is lacking in the present case, The so-called purchaser simultaneously binds himself to re-sell at the same price to the so-called seller. In each instance the goods remained in the possession of the complainants throughout”.
In this regard, it has also been held that the interposition of a third party between the original seller and the final purchaser may lead the Courts to regard such third party as a mere conduit. In the UK case of Furniss v Dawson  AC 474, the Dawson family attempted to avoid capital gains tax by selling their shares in two English trading companies to a special purpose company (SPV) in the Isle of Man in exchange for shares in that company; and immediately thereafter causing the SPV to on-sell the shares to a third party at market-value. In his judgment, Lord Brightman treated the sale as having been concluded between the Dawson family and the third party directly, effectively ignoring the intervention of the SPV. In other words, the SPV was regarded as a mere conduit through which the shares passed:
“When one moves, however, from a single transaction to a series of inter-dependent transactions designed to produce a given result, it is, in my opinion, perfectly legitimate to draw a distinction between the substance and the form of the composite transaction without in any way suggesting that any of the single transactions which make up the whole are other than genuine. This has been the approach of the United States Federal Courts enabling them to develop a doctrine whereby the tax consequences of the composite transaction are dependent on its substance, not its form. I shall not attempt to review the American authorities, nor do I propose a wholesale importation of the American doctrine in all its ramifications into English law. But I do suggest that the distinction between form and substance is one which can usefully be drawn in determining the tax consequences of composite transactions and one which will help to free the Courts from the shackles which have so long been thought to be imposed on them by the Westmister case.”
The principles in Furniss were subsequently summarised by Lord Oliver in Baylis v Gregory at 507:
“As the law currently stands, the essentials emerging from Dawson appear to me to be four in number: (1) that the series of transactions was, at the time when the intermediate transaction was entered into, pre-ordained in order to produce a given result; (2) that the transaction had no other purpose than tax mitigation; (3) that there was at that time no practical likelihood that the pre-planned events would not take place in the order ordained, so that the intermediate transaction was not even contemplated practically as having an independent life; and (4) that the pre-ordained events did in fact take place. In these circumstances the court can be justified in linking the beginning with the end so as to make a single composite whole to which the fiscal results of the single composite whole are to be applied … There is a real and not merely a metaphysical distinction between something that is done as a preparatory step towards a possible but uncertain contemplated future action and something which is done as an integral and interdependent part of a transaction already agreed and, effectively, pre-destined to take place. In the latter case, to link the end to the beginning involves no more than recognising the reality of what is effectively a single operation ab initio. In the former it involves quite a different process, viz that of imputing to the parties, ex post facto, an obligation (either contractual or quasi contractual) which did not exist at the material time but which is to be attributed from the occurrence or juxtaposition of events which subsequently took place. That cannot be extracted from Dawson as it stands not can it be justified by any rational extension of the Ramsay approach.”
Despite rejecting a “belts and braces” adoption of the Furniss principle into our law, a strikingly similar interpretation was relied upon by our Courts in ITC1606 (a case involving an inquiry into section 103(1) of the Act) at headnotes (vi) to (ix):
“That in the present case it appeared from the facts that, although the scheme was made up of three separate steps, each one of which can be seen as genuine, it was clear that they were all parts of one composite transaction and it was also clear that although there were three entities with independent legal personality, two of them took part as members of a single company and the third acted purely as agent for that group in the transaction … H acted purely as agent … That it was clear that H had been brought into the scheme as intermediary exclusively to attempt to reduce tax liability and, accordingly, where H had been brought into the picture purely in order to give effect to the reduction of tax liability, the transaction can be seen in its entirety as being not normal or one at arm’s length … That although the opinion has been expressed that any principle flowing out of the decision in the case of Furniss (Inspector of Taxes) v Dawson  1 ALL ER 530 (HL) should not be adopted by our Courts, common sense holds that when one of the transactions included in a series of transactions has no commercial purpose, and is there exclusively to attempt to obtain a tax advantage, it cannot be seen as a normal transaction between persons acting at arm’s length.“
It is also interesting to note that an equivalent of the Furniss approach is known in US law as the “step approach”. In terms of this approach, each leg of a transaction is regarded in the context of all the other steps in the transaction. However, where an extra step is interposed for no apparent reason, it can be ignored to reveal the true nature of the transaction. In the case of CIR v Court Holding Co 324 U.S. 331 (1945), shareholders in a company owning an apartment complex negotiated the sale of the complex with a third party. However, to avoid tax, the shareholders liquidated the company; had the title to the complex transferred to their individual names; and sold the complex on the terms agreed upon prior to the liquidation. In its decision, the Court held:
“[T]he transaction must be viewed as a whole, and each step, from the commencement of negotiations to the consummation of the sale, is relevant. A sale by one person cannot be transformed for tax purposes into a sale by another using the latter as a conduit through which to pass title.”
Accordingly, the Court ignored the liquidation, and regarded it as an unnecessary step designed to mask or conceal the true intention of the taxpayers, namely to avoid corporate tax. Although the step approach does not necessarily form part of our law, a common principle appears to link all the above cases, irrespective of the name of the doctrine adopted – Furniss principle, step approach, substance over form, or agency.
Applying the above, our Courts could conclude that the real intention of the parties was to assign ownership in the IP from CompanyA to CompanyB, with CompanyC having no interest in the assignment whatsoever. To assist the Court in its reasoning, CompanyC could be found to have acted as an agent of either CompanyA or CompanyB and the role of CompanyC may consequently be ignore.
Individual transactions making up a composite transaction need not be analysed individually. It has been accepted by our Courts that a set of agreements may constitute a “package deal”, which may be analysed as a whole. As Hefer JA in Erf 3183/1 Ladysmith (Pty) Ltd at 953F–954C pointed out:
“Regarded separately and without reference to the others, there is nothing unusual about the terms of the main leases or of any of the other documents except for one remarkable feature. Since the same signatories signed the main leases, the sub-leases and the building contracts simultaneously on behalf of the appellants, the Fund, Pioneer and the contractor respectively, we must infer that they signed each agreement with full knowledge of the terms of the others which were either awaiting their signatures or had already been signed. … Be that as it may be, the agreements cannot be regarded separately: they were all signed simultaneously and were plainly interdependent to the extent that none of them would have been concluded unless all the others were also signed; as the appellants’ attorney conceded, each one must be considered in the context of all the others in order to discover their total effect”
(also see ITC1606, discussed above)
Clearly the series of assignments from CompanyA to CompanyC and from CompanyC to CompanyB constitute a “package deal”: the assignments are concluded on the same day and for the same purchase price; the business of CompanyA is typically transferred to CompanyB in terms of a separate transaction – CompanyB therefore requires ownership of the IP to continue to conduct the business, whereas CompanyC never has any use for the IP.
The next step is to recognise that in terms of the composite transaction, the IP is assigned from CompanyA to CompanyB (via CompanyC). As such, when applying s11(gC)(bb), “the assignor” must be read to refer to CompanyA and “the assignee” to CompanyB. The allowance that may be claimed by CompanyB is thereby limited by s11(gC)(bb) to the lesser of the cost to CompanyA (__R1m__) or the market value of the IP (R200m) – we are back at square one.
It is also worth pointing out that the contra-fiscum rule of interpretation does not assist the taxpayer since s11(gC)(bb) is an anti-avoidance provision. According to Glen Anil Development Corp 1975 (4)( SA 715 (A), anti-avoidance provisions must be interpreted robustly to give effect to the purpose and suppress the mischief against which the section is directed. “The true intention of the legislature is of paramount importance, and, I should say, decisive … [i.e.] defeating tax avoidance schemes.”
Despite being extremely common, the above sale and on-sale arrangement, intended to circumvent connected person anti-avoidance provisions, failed to achieve its goal (on various grounds).